The Emperor's New Saddle
The Ephippium of the Equestrian
of Marcus Aurelius
HELMUT NICKEL
Curatorof Arms and Armor,The MetropolitanMuseum of Art
FOR DIETRICH VON BOTHMER
AMONG THE WORKS OF ART surviving from classi-
cal antiquity, one of the most influential is doubtless
the equestrian statue of Emperor Marcus Aurelius
(reigned A.D. 161-80) (Figure 1). For centuries it
stood in the Lateran, until it was transferred, in
1538, to the Campidoglio by Michelangelo; it sur-
vived largely because it was erroneously believed to
be an effigy of Constantine, the first Christian em-
peror. By a combination of this unjustified attribu-
tion, its own artistic merits, and its conspicuous pres-
ence, it served as a model for the majority of
equestrian statues throughout the entire history of
European art.'
A reduced free copy by Filarete (1400-ca. 1465) is
the earliest known dated small bronze of the Italian
Renaissance (Figure 2). Its plinth bears a presenta-
tion inscription to Piero de'Medici (Filarete'sbenefac-
tor) and the date 1465.2 Although Filarete permitted
himself some artistic liberties by adding an oversize
helmet as a support for the horse's raised foreleg,
and by completing the fragmentary breaststrap of
the original, he took great pains to duplicate the
ephippium, the saddle blanket of the emperor's
mount, with its elaborate border of zigzag cuts,
stepped lappets, and sawtooth patterns.
Half a century later, Francois I had a plaster cast
made of Marcus Aurelius's horse and it was displayed
in one of the courtyards of Fontainebleau, which,
from then on, became known as la Cour du Cheval
Blanc. It remained there until 1626, when it had to
? The Metropolitan Museum of Art 1989
METROPOLITAN MUSEUM JOURNAL24
Cloth:
Statue
IN HONOR OF HIS SEVENTIETH BIRTHDAY
be removed because of weather damage. This horse
was the inspiration for several works of the School of
Fontainebleau-such as an enamel plaque, formerly
in the Lenoir collection, with an equestrian portrait
of Henri II wearing a Roman toga and Diane de Poi-
tiers riding pillion (Figure 3),3 and also the Louvre's
marble relief of Charles IX of France as a Roman im-
perator on horseback; he wears classical parade ar-
mor and strikes a more dramatic pose than the pen-
sive philosopher emperor (Figure 4).4 In spite of the
changes made in the representations of their riders,
the horses in these effigies are faithfully modeled
after Marcus Aurelius's steed, as shown by their me-
ticulously reproduced saddle blankets. This exact
copying even went as far as to include the fragment
of the breaststrap, which in the original is now with-
out a purpose, since the once separately applied cen-
ter piece of the breast harness has been lost.5 Quite
obviously, this particular saddle was considered to be
an authentic piece of Imperial Roman horse equip-
ment; otherwise, it would not have been so carefully
copied.
And as recently as 1951, in a reconstruction of the
lost equestrian statue from the Column of Justinian
at Istanbul, the Byzantine emperor's horse was given
a saddle blanket that bore the pattern of Marcus Au-
relius's ephippium (Figure 5),6 though, according to
late-medieval illustrations of Justinian's statue, he
was originally represented as riding bareback.7
However, the triple-layered construction of the
17
The notes for this articlebegin on page 23.
The Metropolitan Museum of Art
is collaborating with JSTOR to digitize, preserve, and extend access to
Metropolitan Museum Journal ®
www.jstor.org
 IK
'
ri
jl
ppyjISli
i

. . 1 .

1y.
4 .,J
.
' ii .i
,=0
-I'
9
X
^
r
t i d
t E

?
,.
r r-O~ T
'S4/Ns/-y
'k %I

'"
I ?
't I f -I/
h
I.I
1 - -^_ tr

1. Bronze equestrian statue of Marcus Aurelius. Ro-

man, A.D. 2d century. Rome, Campidoglio (photo:
Anderson)

i8
 +11!

'-

/;
I I
? 'i 'i '4 #t i
I

S -

2. Bronze statuette, reduced copy of the statue of Mar- 4. Marble relief of Charles IX on horseback. Paris,
cus Aurelius, by Antonio Averlino, called Filarete Louvre (photo: Arch. Phot. Paris)
(1400-ca. 1465), dated 1465. Dresden, Skulpturen-
sammlung, Staatliche Kunstsammlungen (photo:
Staatliche Kunstsammlungen Dresden)
5. ll.It
Reconstruction of the
Column of Justinian (or 'I- --
- 'r. .

3. Drawing after an enamel representing Henri II and Theodosius). Istanbul A.t

Diane de Poitiers on horseback; formerly Lenoir

collection (after Steinmann)
(after Mamboury) ,'

>;Is j F

- -I . /
tF*J-ari
_ .

CB v - ., .

- ' ,'r 'wa

~f'$il?l

- :1
)
. V .

")e

P-l s
.\.

/
1rB
I? *

.)
I
. )
e? -'S ' ?

17,
=;
.1 L

'. o ..

-
.
i
,-' ^.. 1._
 -' . t&, i m-
Ia.1m m mTimm111111
1 i i
.
.M.
I 1 -1
t?so,,
l
iF~~~~~~~~~~~~~~~~~
J,?1
A
lw~ r
.R~~~~~~~ tpi~
6. Mosaic of the Battle of Alexander and Darius, found
1831 in Pompeii. Naples, National Museum (photo:
Anderson)
ephippium in Marcus Aurelius's statue-with a zig-
zag-edged element on top of one with a border of
stepped lappets, which in turn overlaps one with a
sawtooth border-is actually quite out of the ordi-
nary. Roman saddle blankets of the second century
A.D., as shown in the reliefs of the Trajan and Marcus
Aurelius columns, were mostly simple rectangles of
cloth, sometimes with a heavy fringe at the bottom
edges that hung down below the horses' bellies. The
mounts of the emperors themselves, and those of
their cavalry guard units, sported more elaborate
saddles that had a shorter, zigzag-edged seat cover
on top of the longer fringed blanket.8 This top layer
seems to have been loosely attached to the lower
saddle cloth, because it is sometimes shown draped
over shields hanging on the saddle horns with riders
dismounted; this was probably done to protect the
painted surfaces of the shields in inclement weather.9
Incidentally, the Germanic hostiles are, as a rule, rep-
resented as riding bareback.
20
I r.
I&L.
.
pjC.i~l~ -.''
.. I.-- 4.K-~~~~~~~~~~~,~
I
i/'46,
The two fragmentary equestrian statues from Car-
toceto di Pergola, found in 1946, have ephippia of
similar shape, but simpler construction than that of
Marcus Aurelius.'0 They have single-layer saddle
blankets: one has a sawtooth edge at its bottom and
stepped lappets along its crupper edge, and the
other has stepped lappets all around.
Very similar design arrangements are to be found
in the celebrated mosaic of the Battle of Alexander
and Darius, which was found at Pompeii (Figure 6),"1
where two of the horses on the Persian side are de-
picted with clearly identifiable saddle blankets. On
the collapsing horse of the hapless rider struck by Al-
exander's lance, there is an ephippium with a saw-
tooth border along its bottom and a dagged version
of stepped lappets at its rear edge, while the other
horse-in the center of the composition, in front of
Darius's chariot wheel, where it is held by its dis-
mounted rider, who loyally offers his steed to his
king for a speedier flight-bears a saddle blanket
bordered by stepped lappets. Evidently these saddle
blankets with stepped edges were considered to be
typical for Eastern horsemen, and, indeed, they can
be regularly found on Achaemenian seals and coins,
in the reliefs of Persepolis, on decorated Scythian
sword scabbards, and on Iranian horse rhytons (Fig-
ures 7, 8, 9).12 In 1984, Bernard Goldman coined the
term "the Persian Saddle Blanket" for this peculiar
saddle cloth, although he had to admit that its origins
seem to have been with the steppe nomads in the
North, because the "half merlon," as he calls the
stepped lappet, would be an impractical decoration
for woven fabrics, but eminently suited for felt, the
material favored by the nomads.13
A surviving example of such a saddle blanket of
felt, with stepped lappets along its lower edge, was
found in a state of perfect preservation in one of the
frozen tumuli, known as kurgan V, at Pazyryk in Si-
beria (Figure o1).14 This burial mound of a nomad
prince, from around 300 B.C., also yielded a large felt
tapestry with applique figures of horsemen display-
ing the same saddle blankets with stepped lappets
(Figure 11), and among its further treasures was one
of the earliest known Oriental pile carpets (Figure
12), also with representations of horsemen going
around the border, as if in solemn procession.15 Each
horse bears a saddle blanket edged in stepped lap-
pets.
It seems that these saddle blankets with stepped

7. Seal, Iranian, first millennium B.C.The MetropolitanMuseumof Art, Purchase,

funds from various donors, 1893, 93.17.17

8. Detail of a gold akinakesscabbard,Scythian,ca. 300 B.C.The MetropolitanMuseum

of Art, Rogers Fund, 1930, 30.11.12

9.
Clay horse rhyton, Iranian,
8th-5th century B.C., Teheran
Museum (after Ghirshman)

10. Saddle blanket of felt, found in kurgan V, Pazyryk,
Siberia; probably Sarmatian, ca. 300 B.c. Lenin-
grad, Hermitage Museum (afterJettmar)
edges were an element of material culture shared
among the horsemen of Iranian stock, whether
highly civilized Achaemenian Persians or "barbarian"
nomads, such as Scythians and Sarmatians.16 The no-
mad princes of the Siberian kurgans are thought to
have been Sarmatians. By the first century A.D. Sar-
matian tribes had drifted so far westward that their
outriders made contact with the Romans in the Dan-
ube region.17 These tribes-Alani, Roxolani, Antae,
11. Horseman, detail of a felt tapestry,found in kurgan
V, Pazyryk, Siberia; probably Sarmatian, ca. 300 B.C.
Leningrad, Hermitage Museum (afterJettmar)
and Jazyges-were horse nomads and in warfare
rode as heavy armored cavalry (the Late Roman
heavy cavalry, cataphractariiand clibanarii,was mod-
eled after Sarmatian prototypes).'8 In A.D. 175 Mar-
cus Aurelius succeeded in inflicting a crushing defeat
on the westernmost of these Sarmatians, the Jazyges
of Pannonia, which is now Hungary.19After this mil-
itary success Marcus Aurelius added the honorific
"Sarmaticus"to his name.
From this it seems most likely that Marcus Au-
relius's mount in his equestrian monument is a
captured Sarmatian war steed, which he rides, de-
monstratively saddled in Sarmatian fashion, in cele-
bration of his victory in far-off Pannonia.
APPENDIX
Part of the peace terms with the defeated Jazyges was
that in A.D. 175 they had to contribute 8,000 warriors as
cavalry auxiliaries to the Roman army; 5,500 of these
were sent to Northern Britain attached to the Legio VI
Victrix to fight Pictish would-be invaders.20These Sar-
matians in Britain, incidentally, were not returned to
their homeland after their twenty-yearterm of service
had expired, but were settled in a kibbutzlikemilitary
colony at Bremetennacum, now Ribchester,in present-
day Lancashire, to raise horses for the Roman cavalry
and to guard the coastal area at the mouth of the river
Ribbleagainst Irish pirates. This cuneusveteranorumSar-
matarum at Bremetennacum is still listed in the official
12. Reconstructivedrawing of a pile carpet, found in
kurgan V, Pazyryk, Siberia; probably 4th century
B.C.Leningrad, Hermitage Museum (afterJettmar)
muster roll of the Late Roman army, Notitia Dignitatum,
ca. A.D. 428.21 Considering this late survival, combined
with the facts that Sarmatians were heavy-armored cav-
alry, fought under battle standards in dragon-shape, had
as representations of their tribal god of war a naked
sword thrust in the ground or a platform,22 and finally,
that in A.D. 175 the praefectus of the Legio VI Victrix
was a certain Lucius Artorius Castus,23it is possible that
Marcus Aurelius's victory had another lasting influence,
NOTES
1. Kurt Kluge and Karl Leo Heinrich Lehmann-Hartleben,
Grossbronzen der romischen Kaiserzeit, 3 vols. (Berlin/Leipzig,
1927) II, p. 85; S. A. Callisen,"The EquestrianStatueof Louis
XIV in Dijon and Related Monuments,"The Art Bulletin 23
(1941) pp. 131-140; Harald von Roques de Maumont, Antike
Reiterstandbilder(Berlin, 1958) pp. 56-60, fig. 29; Jocelyn M. C.
Toynbee, Animals in Roman Life and Art (New York, 1973) pp.
172-173; P. Fehl, "The Placement of the Equestrian Statue of
Marcus Aurelius in the Middle Ages," Journal of the WarburgIn-
stitute(1974) pp. 362-368; Licia Borelli Vlad, Giulia Fogolari,
and Anna Guidi Toniato, "The stylisticproblemof the Horses
of San Marco" in The Horses of San Marco, Venice(London, 1979)
pp. 15-41, fig. 49; and Die Pferde von San Marco in Venedig,exh.
cat. (Berlin, 1982) cat. nos. 177, 178, 180-184.
2. Wolfgang von Oettingen, Antonio Averlino, genannt Filarete,
in Beitrage zur Kunstgeschichte6 (Leipzig, 1888) pp. 13-14; and
Joachim Menzhausen, "MarcusAurelius on Horseback,"The
Splendorof Dresden,exh. cat. (Washington/NewYork/SanFran-
cisco, 1978) no. 47, p. 85, ill. The statuettecame to the Kunst-
kammerof PrinceElectorChristianI of Saxony(r. 1586-91) as
a present from GuglielmoGonzaga,Duke of Mantua,in the late
16th century,and is now inv. no. H 155/37 in the Skulpturen-
sammlung,Dresden.
3. Ernst Steinmann, "Die Zerstorungder Konigsdenkmaler
in Paris," Monatshefteder Kunstwissenschaft(1917) pp. 337-380,
pl. 52, fig. 9; and Callisen,"EquestrianStatueof Louis XIV,"p.
136, fig. 8.
4. Michele Beaulieu, "Le Roi CharlesIX a Cheval,"Fontaine-
bleau: Art in France, 1528-1610, 2 vols., exh. cat. (Ottawa, 1973)
II, no. 540, p. 117, cover ill.; Musee du Louvre, inv. R. F. 1474,
attributedto GermainPilon (1528-90). The reversalof the po-
sition of the horse might be the result of its being copied after
a print.
5. The now-lostcenter section of the breaststrapand the me-
dallions, also lost, on the horse'sheadstallwere possiblyenam-
eled (as in Filarete'sreduction)and were therefore made sepa-
rately.
6. Ernest Mamboury, Istanbul Touristique(Istanbul, 1951) p.
261; and Guido Perocco,"The Horses of San Marcoin Venice,"
in The Horses of San Marco, pp. 49-79, fig. 81; see also "Die
Pferde von San Marco in Venedig," in Die Pferde von San Marco,
PP- 55-72, fig. 36.
this time on medieval literature, by contributing to the
development of legends about King Arthur.24
ACKNOWLEDGMENTS
For much appreciated help and generously shared in-
formation I would like to thank my friends and col-
leagues Elfriede R. Knauer, Joan R. Mertens, Stuart W.
Pyhrr, and George Szabo.
7. Phyllis WilliamsLehmann, "Theodosius or Justinian?A
RenaissanceDrawingof a ByzantineRider,"TheArtBulletin41
(Mar. 1959).
8. Lino Rossi, Trajan'sColumn and the Dacian Wars (Ithaca,
N.Y., 1971) p. 135, figs. 7, 8; p. 142, fig. 19; p. 149, fig. 31; p.
151, fig. 32; p. 157, figs. 42, 43; p. 159, fig. 44; p. 178, figs. 78,
79; and C. Caprino et al., La Colonna di Marco Aurelio (Rome,
1955) pl. IV, fig. 9 (III); pl. xvIII, figs. 36, 37 (xxvi, xxviI,
xxvIIi); pl. xxxII, fig. 65 (LI); pi. XXXVI,figs. 71, 72 (LVI); pl.
XLVI,figs. 91, 92 (LXXIV,LXXV);pl. L, fig. 99 (LXXIX);pl. LV,fig.
109 (XCI);pl. LVII, fig. 114 (xcv); pl. LVIII, fig. 115 (xcv); and
pl. LXIV, fig. 128 (cvIII); see also H. Russell Robinson, The Ar-
mour of Imperial Rome (New York, 1975) chap. 7, "Horse Ar-
mour," pp. 190-196; and Paul Holder and Peter Connolly, "Co-
hortes Equitataefrom Augustusto Hadrian,"MilitaryIllustrated:
Past & Present 13 (June-July 1988) pp. 21-32, ill.
Among the sixteen monuments with saddle blankets illus-
trated in Roques de Maumont,Reiterstandbilder,
only the ephip-
pium of Marcus Aurelius's mount (fig. 29) shows stepped lap-
pets; on one statuette from Pompeii (fig. 11), perhaps
representing Alexander the Great, the ephippium has a crenel-
lated fringe; on three other statues (figs. 10, 30, 50) the saddle
blanketsare animalskins with the paws still attached.
9. Rossi, Trajan'sColumn, p. 157, figs. 42, 43; p. 159, fig. 44.
o1.Bronzi Dorati da Cartoceto(Cantini, 1987) pls. 34, 39, 40;
and in Die Pferde von San Marco, pp. 188-189, cat. nos. 91, 92.
11. Unfortunately,there are great lacunae in the mosaic.
Therefore, neither Alexander's saddle nor any of those of the
other Greek horsemen have been preserved; they could have
offered a comparison.
12. Hans Henning von der Osten, "The Ancient Seals from
the Near East in the Metropolitan Museum," MMAB 13 (1931)
2, no. 14; idem, Die Weltder Perser (Stuttgart, 1956) pls. 18, 86;
Roman Ghirshman, The Arts of AncientIran: FromIts Origin to the
Time of Alexander the Great (New York, 1964) p. 261, fig. 345; p.
288, fig. 346; p. 289, fig. 347; p. 291, fig. 348; and idem, "La
Selle en Iran," Iranica Antiqua 10 (1973); pp. 94-107, figs. 2, 3,
pl. L.
13. Bernard Goldman, "The Persian Saddle Blanket," Studia
Iranica 13 (1984) pp. 7-18, pls. I-vI; and Elfriede R. Knauer,
"Gleanings," Studia Iranica 15 (1986) pp. 265-266, pl. i.
Knauer draws attention to the statue of Marcus Aurelius, in
23
response to Goldman, who had stated that the "Persian Saddle
Blanket," formerly a status symbol of nobility, had disappeared
in the second century B.C.
14. Karl Jettmar, Art of the Steppes (New York, 1964) p. 13,
pl. 17 (saddle blanket), p. 115, pl. 18 (tapestry), pp. 118-121,
pl. 19, fig. 103 (carpet); Tamara Talbot-Rice, AncientArts of Cen-
tral Asia (New York, 1965) p. 35, fig. 26 (carpet), pp. 38-39, fig.
30 (tapestry); E. D. Phillips, The Royal Hordes: Nomad Peoples of
the Steppes(London/New York, 1965) pp. 83-85, fig. 92 (carpet),
93 (tapestry); Mikhail P. Gryaznov, The Ancient Civilization of
Southern Siberia: An ArchaeologicalAdventure (New York, 1969)
pi. 118 (saddle blanket), pls. 134-135 (carpet), pl. 137 (tapes-
try); and Ghirshman, Arts of AncientIran, pp. 360, 361, figs. 466,
467 (carpet).
It is interesting that the surviving saddle blanket has a frag-
ment of the breaststrap remaining at exactly the same spot as in
Marcus Aurelius's statue.
15. The carpet measures 200 by 190 cm. and has 36 knots per
cm. (215 knots to the square inch). In the outer bordures (grif-
fin medallions, horsemen, and star bursts) are markings sug-
gesting starting points for a dicing game.
16. In Die Pferde von San Marco, cat. no. 143, there is a bronze
statuette of a centaur, described there as workshop of Riccio, ca.
1500, which has a caparison with a triple border of stepped lap-
pets and sawtooth patterns copied from Marcus Aurelius's
ephippium. Cat. no. 43 is a terra-cotta relief plaque of the sec-
ond half of the 6th century B.C., from Southwestern Turkey,
which shows a horseman pursuing a griffin. Here the saddle
blanket is painted on with a large sawtooth pattern (N. Thomas,
"Recent Acquisitions by Birmingham City Museum," Archaeolog-
ical Reportsfor 1964-1965 [1965] pp. 64ff.). Cat. no. 140 is a
drawing by Polidoro da Caravaggio (1490/1500-48?) after the
Ammendola sarcophagus, ca. A.D. 180 in the National Museum,
Rome, which shows the hindquarters of a horse with a saddle
blanket edged in a sawtooth cut. This horse, ridden by a Roman
officer slaying a barbarian, was possibly meant to be a captured
steed; the same might be true of the saddle horses wearing seat
covers with zigzag edges that are found among the cavalry cam-
paigning in the Balkans against Dacians and Sarmatians in the
friezes of Trajan's Column. For the Ammendola sarcophagus,
see Bernard Andreae, MotivgeschichtlicheUntersuchungenzu den
romischenSchlachtsarkophagen(Berlin, 1956) pl. 4.
One of the Three Wise Men on horseback in the sculptural
decoration of the so-called Trivulzio candelabrum in the cathe-
dral of Milan has an unusual saddle blanket with stepped lap-
pets. Apparently, the 12th-century artist was still aware of a tra-
dition that saddles of this type were "Eastern," and therefore
appropriate for the Three Kings from the Orient. For the
Milan candelabrum, see Otto Homburger, Der Trivulzio-Kande-
laber (Zurich, 1949) pl. 45.
24
17. Tadeusz Sulimirski, "The Forgotten Sarmatians: A once-
mighty folk scattered among the nations," chap. 12 in Vanished
Civilizations(London/New York, 1963) pp. 279-298; and idem,
The Sarmatians(New York, 1970).
18. Ortwin Gamber, "Dakische und sarmatische Waffen auf
den Reliefs der Trajanssaule," Jahrbuch der Kunsthistorischen
Sammlungenzu Wien 60 (1964) pp. 74ff.; J. W. Eadie, "The De-
velopment of Roman Mailed Cavalry," The Journal of Ro-
man Studies 57 (1967) pp. 161-173; and Ortwin Gamber,
"Kataphrakten, Clibanarier, Normannenreiter," Jahrbuch der
Kunsthistorischen.SammlungenWien 64 (1968) pp. 7ff.
19. Earnest Cary, Dio's Roman History (London/New York,
1927) epitome of Book LXXII, 22.1, pp. 35-36; Roques de
Maumont, Reiterstandbilder,p. 58, rejects Max Wegner's theory
in Herrscherbildnissein antoninischerZeit (Berlin, 1939) p. 114,
that the statue should have been erected when Marcus Aurelius
added the title "Armenicus" to his name in A.D. 164. Instead, he
suggests that it might have been shortly after A.D. 147, when
Marcus Aurelius became co-regent.
Armenia, though, is within the range of distribution of the
"Persian Saddle Blanket"; see Burchard Brentjes, Drei ahrtau-
sende Armenien (Leipzig, 1973) pl. 5, where he illustrates the
horseman rhyton found at Erebuni, Armenia, in 1968. The
ephippia of the Cartoceto statues, which are thought to date
from the early first century B.C., might originate from Asia Mi-
nor as they are too early to be Balkan booty; but they might be
diplomatic presents from somewhere in the steppes.
Richard Brilliant, Roman Art from the Republic to Constantine
(London, 1974) p. 115, fig. 11.33, dates it at "about A.D. 176" in
the caption.
20. Cary, Dio's History, Book LXXII, 22.1, 2, pp. 35-36; Suli-
mirski, "Forgotten Sarmatians," p. 293; and idem, Sarmatians,
pp. 175-176, fig. 65.
21. Otto Seeck, Notitia Dignitatum (Berlin, 1876) OC. XL, 54,
p. 212; and I. A. Richmond, "The Sarmatae, Bremetennacum
Veteranorum and the Regio Bremetennaciensis," TheJournal of
Roman Studies 35 (1945) pp. 16-29.
22. Herodotus VI, 62, and Ammianus Marcellinus XXXI,
2.23.
23. Kemp Malone, "Artorius,"ModernPhilology 22 (1924) pp.
367ff.
24. Helmut Nickel, "Wer waren Konig Artus' Ritter? Uber die
geschichtliche Grundlage der Artussagen," Waffen-und Kostum-
kunde (1975) Heft 1, pp. 1-28; idem, "The Dawn of Chivalry,"
From the Lands of the Scythians:Ancient Treasuresfrom theMuseums
of the U.S.S.R., 3000 B.C.-100 B.C., exh. cat. (New York, 1975)
pp. 150-152; and C. Scott Littleton and Ann C. Thomas, "The
Sarmatian Connection: New Light on the Origin of the Arthu-
rian and Holy Grail Legends," Folklore91 (1978) pp. 513-527.